core.models.uma.nn.mole

core.models.uma.nn.mole#

Copyright (c) Meta Platforms, Inc. and affiliates.

This source code is licensed under the MIT license found in the LICENSE file in the root directory of this source tree.

Attributes#

fairchem_cpp_found

fairchem_cpp_found

Classes#

MOLEGlobals

MOLEDGL

Base class for all neural network modules.

MOLE

Base class for all neural network modules.

Functions#

_softmax(x)

_pnorm(x)

norm_str_to_fn(act)

init_linear(num_experts, use_bias, out_features, ...)

Module Contents#

core.models.uma.nn.mole.fairchem_cpp_found = False#
core.models.uma.nn.mole.fairchem_cpp_found = True#
core.models.uma.nn.mole._softmax(x)#
core.models.uma.nn.mole._pnorm(x)#
core.models.uma.nn.mole.norm_str_to_fn(act)#
class core.models.uma.nn.mole.MOLEGlobals#
expert_mixing_coefficients: torch.Tensor#
mole_sizes: torch.Tensor#
core.models.uma.nn.mole.init_linear(num_experts, use_bias, out_features, in_features)#
class core.models.uma.nn.mole.MOLEDGL(num_experts, in_features, out_features, global_mole_tensors, bias: bool)#

Bases: torch.nn.Module

Base class for all neural network modules.

Your models should also subclass this class.

Modules can also contain other Modules, allowing them to be nested in a tree structure. You can assign the submodules as regular attributes:

import torch.nn as nn
import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.conv1 = nn.Conv2d(1, 20, 5)
        self.conv2 = nn.Conv2d(20, 20, 5)

    def forward(self, x):
        x = F.relu(self.conv1(x))
        return F.relu(self.conv2(x))

Submodules assigned in this way will be registered, and will also have their parameters converted when you call to(), etc.

Note

As per the example above, an __init__() call to the parent class must be made before assignment on the child.

Variables:

training (bool) – Boolean represents whether this module is in training or evaluation mode.

num_experts#
in_features#
out_features#
global_mole_tensors#
forward(x)#
class core.models.uma.nn.mole.MOLE(num_experts, in_features, out_features, global_mole_tensors: MOLEGlobals, bias: bool)#

Bases: torch.nn.Module

Base class for all neural network modules.

Your models should also subclass this class.

Modules can also contain other Modules, allowing them to be nested in a tree structure. You can assign the submodules as regular attributes:

import torch.nn as nn
import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.conv1 = nn.Conv2d(1, 20, 5)
        self.conv2 = nn.Conv2d(20, 20, 5)

    def forward(self, x):
        x = F.relu(self.conv1(x))
        return F.relu(self.conv2(x))

Submodules assigned in this way will be registered, and will also have their parameters converted when you call to(), etc.

Note

As per the example above, an __init__() call to the parent class must be made before assignment on the child.

Variables:

training (bool) – Boolean represents whether this module is in training or evaluation mode.

num_experts#
in_features#
out_features#
global_mole_tensors#
merged_linear_layer()#
forward(x)#
Contents